Navigating the tradeoffs between environmental DNA and conventional field surveys for improved amphibian monitoring

The need for efficient, accurate biodiversity monitoring is growing, especially for globally imperiled taxa, such as amphibians. Environmental DNA (eDNA) analysis holds enormous potential for enhancing monitoring programs, but as this tool is increasingly adopted, it is imperative for users to understand its potential benefits and shortcomings. We conducted a comparative study to evaluate the efficacy of two eDNA methodologies (quantitative (q)PCR and metabarcoding) and conventional field sampling approaches (seining, dipnetting, and visual encounter surveys) in a system of 20 ponds containing six different amphibian species. Using an occupancy modeling framework, we estimated differences in detection sensitivity across methods, with a focus on how eDNA survey design could be further optimized. Overall, both metabarcoding and qPCR were competitive with or improved upon conventional methods. Specifically, qPCR (a species-specific approach) was the most effective technique for detecting two rare species, the California tiger salamander (Ambystoma californiense) and California red-legged frog (Rana draytonii), with a detection probability of greater than 0.80 per survey. Metabarcoding (a community approach), estimated amphibian diversity with comparable rates to field techniques on average, and detected an additional 41 vertebrate taxa. However, for two abundant species (western toads, Anaxyrus boreas, and Pacific chorus frogs, Pseudacris regilla), field techniques outperformed metabarcoding, especially towards the latter half of the summer. Our results indicate that eDNA approaches would be most effective when paired with visual encounter surveys to detect terrestrial life-stages, and that more optimization, specifically primer choice and validation, is needed. By comparing methods across a diverse set of ponds and species, we provide guidance for future studies incorporating eDNA approaches into amphibian monitoring.